44fc09ac64d99b7426d53102d4aa972509f50644 angie Wed Apr 7 19:47:33 2021 -0700 Oops, overdue adding this work done over a year ago to git. diff --git src/hg/makeDb/doc/hg38/grcH38P13.txt src/hg/makeDb/doc/hg38/grcH38P13.txt new file mode 100644 index 0000000..6dbedbc --- /dev/null +++ src/hg/makeDb/doc/hg38/grcH38P13.txt @@ -0,0 +1,383 @@ +# for emacs: -*- mode: sh; -*- + +############################################################################## +# GRCh38 patch 13 build: build basic tracks on separate database for new +# sequences only (relative to hg38). +############################################################################## + +############################################################################## +# download or rather ln -s the patch release files (DONE - 2020-02-26 - Angie) + + # Note: newer assemblies use refseq releases instead of genbank, but hg38 uses genbank + # so continue with that when building patches. + mkdir -p /hive/data/genomes/grcH38P13/genbank + cd /hive/data/genomes/grcH38P13/genbank + + # Releases have already been downloaded to /hive/data/outside/ncbi/genomes/. + ln -s /hive/data/outside/ncbi/genomes/genbank/vertebrate_mammalian/Homo_sapiens/all_assembly_versions/GCA_000001405.28_GRCh38.p13/* . + + +############################################################################## +# Set up fasta and agp with UCSC names (DONE - 2020-02-26 - Angie) + + mkdir /hive/data/genomes/grcH38P13/ucsc + cd /hive/data/genomes/grcH38P13/ucsc + + # identify sequences not in existing genome db + faCount ../genbank/GCA_000001405.28_GRCh38.p13_genomic.fna.gz \ + > faCount.GRCh38.p13.txt + ~/kent/src/hg/makeDb/doc/hg38/scanAssemblyReport.pl \ + /hive/data/genomes/hg38/chrom.sizes \ + faCount.GRCh38.p13.txt ../genbank/GCA_000001405.28_GRCh38.p13_assembly_report.txt \ + | grep -w new > new.sequences.list + wc -l new.sequences.list +#45 new.sequences.list + + # Extract UCSC-named FASTA for the new sequences + cut -f3 new.sequences.list > extract.new.list + awk '{printf "s/%s/%s/; ", $3,$1}' new.sequences.list > genbankToUCSC.sed + faSomeRecords ../genbank/GCA_000001405.28_GRCh38.p13_genomic.fna.gz extract.new.list stdout \ + | sed -e 's/ .*//;' \ + | sed -f genbankToUCSC.sed \ + | gzip -c > grcH38P13.fa.gz + faSize grcH38P13.fa.gz +#14769668 bases (0 N's 14769668 real 8821498 upper 5948170 lower) in 45 sequences in 1 files +#Total size: mean 328214.8 sd 802643.1 min 7309 (chr13_ML143363v1_fix) max 5500449 (chr15_ML143371v1_fix) median 192531 + + # Compare faSize results for whole GCA_000001405.28_GRCh38.p13_genomic.fna.gz + # vs. concatenation of hg38 fasta and grcH38P13.fa.gz: + faSize ../genbank/GCA_000001405.28_GRCh38.p13_genomic.fna.gz +#3272116950 bases (161368694 N's 3110748256 real 1928965269 upper 1181782987 lower) in 640 sequences in 1 files +#Total size: mean 5112682.7 sd 26768936.7 min 970 (KI270394.1) max 248956422 (CM000663.2) median 166743 + + twoBitToFa /hive/data/genomes/hg38/hg38.2bit stdout \ + | faSize grcH38P13.fa.gz stdin +#3272116950 bases (161368694 N's 3110748256 real 1491934681 upper 1618813575 lower) in 640 sequences in 2 files +#Total size: mean 5112682.7 sd 26768936.7 min 970 (chrUn_KI270394v1) max 248956422 (chr1) median 166743 + # Good, everything in GCA_000001405.28_GRCh38.p13_genomic.fna.gz is accounted for + # between hg38.2bit and grcH38P13.fa.gz + + # Make UCSC-named AGP: + zcat ../genbank/GCA_000001405.28_GRCh38.p13_assembly_structure/PATCHES/alt_scaffolds/AGP/alt.scaf.agp.gz \ + | grep -Fwf extract.new.list \ + | sed -f genbankToUCSC.sed > grcH38P13.agp + + # construct 2bit file: + cd /hive/data/genomes/grcH38P13 + faToTwoBit ucsc/grcH38P13.fa.gz grcH38P13.unmasked.2bit + twoBitInfo grcH38P13.unmasked.2bit stdout | sort -k2nr > chrom.sizes + # take a look at chrom.sizes to verify it looks OK. + + # Make sure AGP and FASTA/2bit agree: + checkAgpAndFa ucsc/grcH38P13.agp grcH38P13.unmasked.2bit | tail -1 +#All AGP and FASTA entries agree - both files are valid + + +############################################################################## +# establish config.ra file (DONE - Angie - 2020-02-26) + # arguments here are: <db> <clade> <trackDbDir> <assembly_report.txt> + cd /hive/data/genomes/grcH38P13 + # Must make photoReference.txt first -- copy from hg38 + cp /hive/data/genomes/hg38/photoReference.txt . + $HOME/kent/src/hg/utils/automation/prepConfig.pl grcH38P13 haplotypes \ + GRCh38.p13 genbank/*_assembly_report.txt > grcH38P13.config.ra + # Edit grcH38P13.config.ra to avoid confusion with actual hg38 +assemblyDate Feb. 2019 p13 +orderKey 2001 + + sed -e 's/^/#/' grcH38P13.config.ra +## config parameters for makeGenomeDb.pl: +#db grcH38P13 +#clade haplotypes +#scientificName Homo sapiens +#commonName Human +#assemblyDate Feb. 2019 p13 +#assemblyLabel Genome Reference Consortium +#assemblyShortLabel GRCh38.p13 +#orderKey 2001 +## mitochondrial sequence included in refseq release +## mitoAcc J01415.2 +#mitoAcc none +#fastaFiles /hive/data/genomes/grcH38P13/ucsc/*.fa.gz +#agpFiles /hive/data/genomes/grcH38P13/ucsc/*.agp +## qualFiles none +#dbDbSpeciesDir GRCh38.p13 +#photoCreditURL http://www.cbse.ucsc.edu/ +#photoCreditName Graphic courtesy of CBSE +#ncbiGenomeId 51 +#ncbiAssemblyId 2334371 +#ncbiAssemblyName GRCh38.p13 +#ncbiBioProject 31257 +#ncbiBioSample notFound +#genBankAccessionID GCA_000001405.28 +#taxId 9606 + + +############################################################################## +# Initial database build (DONE - 2020-02-26 - Angie) + + cd /hive/data/genomes/grcH38P13 + # AGP and unmasked.2bit are already built and checked, so start at the db step: + mkdir jkStuff + $HOME/kent/src/hg/utils/automation/makeGenomeDb.pl grcH38P13.config.ra -debug +#HgStepManager: executing from step 'seq' through step 'trackDb'. +#HgStepManager: executing step 'seq' Wed Feb 26 16:55:27 2020. +#seq: looks like this was run successfully already (/cluster/data/grcH38P13/chrom.sizes exists). Either run with -continue agp or some later step, or move aside/remove /cluster/data/grcH38P13/chrom.sizes and run again. + + # Make chromInfo.tab. + mkdir -p bed/chromInfo + awk '{print $1 "\t" $2 "\t/gbdb/grcH38P13/grcH38P13.2bit";}' chrom.sizes \ + > bed/chromInfo/chromInfo.tab + # Make a link to the .agp file where makeGenomeDb.pl expects to find it. + ln -s ucsc/grcH38P13.agp . + + # Skip 'seq' and 'agp' steps because those were done above. + time ($HOME/kent/src/hg/utils/automation/makeGenomeDb.pl \ + -workhorse=hgwdev -dbHost=hgwdev -fileServer=hgwdev \ + -continue=db grcH38P13.config.ra) > db.log 2>&1 & + tail -f db.log +#real 0m22.790s + + # Ignore all the "NOTES -- STUFF THAT YOU WILL HAVE TO DO --" stuff because this is + # going to be folded into hg38. + # Now the gold, gap and gc5BaseBw tracks are built. + + +############################################################################# +# RepeatMasker (DONE - 2020-02-27 - Angie) + mkdir /hive/data/genomes/grcH38P13/bed/repeatMasker + cd /hive/data/genomes/grcH38P13/bed/repeatMasker + time ($HOME/kent/src/hg/utils/automation/doRepeatMasker.pl -buildDir=`pwd` \ + -bigClusterHub=ku -dbHost=hgwdev -workhorse=hgwdev \ + -smallClusterHub=ku grcH38P13) > do.log 2>&1 & + tail -f do.log +#real 58m13.702s + + egrep "bases|Total|masked" faSize.rmsk.txt \ + | fold -w 95 -s | sed -e 's/^/# /;' +# 14769668 bases (0 N's 14769668 real 6559203 upper 8210465 lower) in 45 sequences in 1 files +# Total size: mean 328214.8 sd 802643.1 min 7309 (chr13_ML143363v1_fix) max 5500449 +# (chr15_ML143371v1_fix) median 192531 +# %55.59 masked total, %55.59 masked real + + egrep -i "versi|relea" do.log | sed -e 's/^/# /;' +# RepeatMasker version development-$Id: RepeatMasker,v 1.332 2017/04/17 19:01:11 rhubley Exp $ +# grep version of RepeatMasker$ /hive/data/staging/data/RepeatMasker/RepeatMasker +# # February 01 2017 (open-4-0-8) 1.332 version of RepeatMasker +# grep RELEASE /hive/data/staging/data/RepeatMasker/Libraries/RepeatMaskerLib.embl +# CC Dfam_Consensus RELEASE 20181026; * +# CC RepBase RELEASE 20181026; * + + featureBits -countGaps grcH38P13 rmsk +#8210465 bases of 14769668 (55.590%) in intersection + + +########################################################################## +# running simple repeat (DONE - 2020-02-27 - Angie) + + mkdir /hive/data/genomes/grcH38P13/bed/simpleRepeat + cd /hive/data/genomes/grcH38P13/bed/simpleRepeat + # using trf409 6 here like hg38 + time ($HOME/kent/src/hg/utils/automation/doSimpleRepeat.pl -buildDir=`pwd` \ + -dbHost=hgwdev -workhorse=hgwdev -bigClusterHub=ku -smallClusterHub=ku \ + -trf409 6 grcH38P13) > do.log 2>&1 & +#real 6m49.141s + + cat fb.simpleRepeat +#1549326 bases of 14769668 (10.490%) in intersection + + # adding this trfMask to .rmsk.2bit + cd /hive/data/genomes/grcH38P13 + twoBitMask grcH38P13.rmsk.2bit \ + -add bed/simpleRepeat/trfMask.bed grcH38P13.2bit + # you can safely ignore the warning about fields >= 13 + + twoBitToFa grcH38P13.2bit stdout | faSize stdin > faSize.grcH38P13.2bit.txt + egrep "bases|Total|masked" faSize.grcH38P13.2bit.txt \ + | fold -w 95 -s | sed -e 's/^/# /;' +# 14769668 bases (0 N's 14769668 real 6528429 upper 8241239 lower) in 45 sequences in 1 files +# Total size: mean 328214.8 sd 802643.1 min 7309 (chr13_ML143363v1_fix) max 5500449 +# (chr15_ML143371v1_fix) median 192531 +# %55.80 masked total, %55.80 masked real + + # reset the symlink + ln -sf `pwd`/grcH38P13.2bit /gbdb/grcH38P13/grcH38P13.2bit + + +########################################################################## +## WINDOWMASKER (DONE - 2020-02-27 - Angie) + + mkdir /hive/data/genomes/grcH38P13/bed/windowMasker + cd /hive/data/genomes/grcH38P13/bed/windowMasker + time ($HOME/kent/src/hg/utils/automation/doWindowMasker.pl -buildDir=`pwd` \ + -workhorse=hgwdev -dbHost=hgwdev grcH38P13) > do.log 2>&1 +# *** All done ! - Elapsed time: 0m47s + + featureBits -countGaps grcH38P13 rmsk windowmaskerSdust \ + > fb.grcH38P13.rmsk.windowmaskerSdust.txt 2>&1 + cat fb.grcH38P13.rmsk.windowmaskerSdust.txt +#3364201 bases of 14769668 (22.778%) in intersection + + # Masking statistics + egrep "bases|Total|masked" faSize.grcH38P13.cleanWMSdust.txt \ + | fold -w 95 -s | sed -e 's/^/# /;' +# 14769668 bases (0 N's 14769668 real 10765455 upper 4004213 lower) in 45 sequences in 1 files +# Total size: mean 328214.8 sd 802643.1 min 7309 (chr13_ML143363v1_fix) max 5500449 +# (chr15_ML143371v1_fix) median 192531 +# %27.11 masked total, %27.11 masked real + + +############################################################################# +# cytoBandIdeo - (DONE - 2020-02-27 - Angie) + mkdir /hive/data/genomes/grcH38P13/bed/cytoBand + cd /hive/data/genomes/grcH38P13/bed/cytoBand + makeCytoBandIdeo.csh grcH38P13 + + +############################################################################# +# cpgIslands - (DONE - 2020-02-27 - Angie) + mkdir /hive/data/genomes/grcH38P13/bed/cpgIslands + cd /hive/data/genomes/grcH38P13/bed/cpgIslands + time ($HOME/kent/src/hg/utils/automation/doCpgIslands.pl -dbHost=hgwdev \ + -workhorse=hgwdev -smallClusterHub=ku -bigClusterHub=ku grcH38P13) > do.log 2>&1 & +# *** All done ! Elapsed time: 1m10s + + cat fb.grcH38P13.cpgIslandExt.txt +#232897 bases of 14769668 (1.577%) in intersection + + +############################################################################## +# genscan - (DONE - 2020-02-27 - Angie) + mkdir /hive/data/genomes/grcH38P13/bed/genscan + cd /hive/data/genomes/grcH38P13/bed/genscan + time ($HOME/kent/src/hg/utils/automation/doGenscan.pl -buildDir=`pwd` \ + -workhorse=hgwdev -dbHost=hgwdev -bigClusterHub=ku grcH38P13) > do.log 2>&1 & +# *** All done ! Elapsed time: 3m11s + + cat fb.grcH38P13.genscan.txt +#478205 bases of 14769668 (3.238%) in intersection + + cat fb.grcH38P13.genscanSubopt.txt +#442575 bases of 14769668 (2.997%) in intersection + + +############################################################################# +# augustus gene track (DONE - 2020-02-27 - Angie) + + mkdir /hive/data/genomes/grcH38P13/bed/augustus + cd /hive/data/genomes/grcH38P13/bed/augustus + time ($HOME/kent/src/hg/utils/automation/doAugustus.pl -buildDir=`pwd` -bigClusterHub=ku \ + -species=human -dbHost=hgwdev -workhorse=hgwdev grcH38P13) > do.log 2>&1 & +# *** All done ! Elapsed time: 30m26s + + cat fb.grcH38P13.augustusGene.txt +#418904 bases of 14769668 (2.836%) in intersection + + +############################################################################## +# Download files (DONE - 2020-02-27 - Angie) + cd /hive/data/genomes/grcH38P13 + time ($HOME/kent/src/hg/utils/automation/makeDownloads.pl -noChromFiles \ + -workhorse=hgwdev grcH38P13) > downloads.log 2>&1 & +# *** All done! +#real 0m20.430s + + +############################################################################## +# PREPARE LINEAGE SPECIFIC REPEAT FILES FOR LASTZ (DONE - 2020-02-27 - Angie) + mkdir /hive/data/genomes/grcH38P13/bed/linSpecRep + cd /hive/data/genomes/grcH38P13/bed/linSpecRep + # create individual .out files from the master record in ../repeatMasker + # cluster job is perhaps overkill for just the patches, but it works + mkdir splitOut + cat <<'_EOF_' > split.csh +#!/bin/csh -fe +set C = $1 +head -3 ../repeatMasker/grcH38P13.sorted.fa.out > splitOut/${C}.out +grep "${C} " ../repeatMasker/grcH38P13.sorted.fa.out >> splitOut/${C}.out +_EOF_ + chmod +x split.csh + + cat << '_EOF_' > template +#LOOP +split.csh $(root1) {check out line+ splitOut/$(root1).out} +#ENDLOOP +_EOF_ + # small ones first: + cut -f1 ../../chrom.sizes | tac > chrom.list + gensub2 chrom.list single template jobList + para make jobList + para time +#Completed: 45 of 45 jobs +#CPU time in finished jobs: 1s 0.01m 0.00h 0.00d 0.000 y +#IO & Wait Time: 132s 2.21m 0.04h 0.00d 0.000 y +#Average job time: 3s 0.05m 0.00h 0.00d +#Longest finished job: 5s 0.08m 0.00h 0.00d +#Submission to last job: 20s 0.33m 0.01h 0.00d + + # now, we can date and process each of those .out files + # constructing the humanSpecific set of repeats + # this means repeats found in human, and not in others + # using mouse here for 'others' is good enough, a variety + # of other species could be used (rat dog cow) where they all + # produce the same result + mkdir dateRepeats + cd dateRepeats + cat << '_EOF_' > mkLSR +#!/bin/bash +set -beEu -o pipefail +seq=$1 +rm -f $seq.out_mus-musculus +ln -sf ../splitOut/$seq.out . +/scratch/data/RepeatMasker/DateRepeats $seq.out -query human -comp 'mus musculus' +rm $seq.out +mkdir -p ../humanSpecific +/cluster/bin/scripts/extractRepeats 1 $seq.out_mus-musculus \ + > ../humanSpecific/$seq.out.spec +_EOF_ + chmod +x mkLSR + + cat << '_EOF_' > template +#LOOP +./mkLSR $(path1) {check out line+ ../humanSpecific/$(path1).out.spec} +#ENDLOOP +_EOF_ + + gensub2 ../chrom.list single template jobList + para make jobList + para time +#Completed: 45 of 45 jobs +#CPU time in finished jobs: 893s 14.88m 0.25h 0.01d 0.000 y +#IO & Wait Time: 95s 1.59m 0.03h 0.00d 0.000 y +#Average job time: 22s 0.37m 0.01h 0.00d +#Longest finished job: 26s 0.43m 0.01h 0.00d +#Submission to last job: 59s 0.98m 0.02h 0.00d + + # We also need the nibs for blastz runs with lineage specific repeats + mkdir /hive/data/genomes/grcH38P13/bed/nibs + cd /hive/data/genomes/grcH38P13/bed/nibs + cut -f1 ../../chrom.sizes | while read C +do + twoBitToFa -seq=${C} ../../grcH38P13.2bit stdout \ + | faToNib -softMask stdin ${C}.nib + echo "${C} done" +done + + # verify nothing lost + cat ../../chrom.sizes \ + | awk '{printf "nibFrag -masked %s.nib 0 %d + stdout\n", $1, $2}' \ + | sh | faSize stdin +#14769668 bases (0 N's 14769668 real 6528429 upper 8241239 lower) in 45 sequences in 1 files +#Total size: mean 328214.8 sd 802643.1 min 7309 (chr13_ML143363v1_fix.nib:0-7309) max 5500449 (chr15_ML143371v1_fix.nib:0-5500449) median 192531 +#N count: mean 0.0 sd 0.0 +#U count: mean 145076.2 sd 355830.6 +#L count: mean 183138.6 sd 450022.5 +#%55.80 masked total, %55.80 masked real + + mkdir -p /hive/data/staging/data/grcH38P13/nib + rsync -a --progress ./ /hive/data/staging/data/grcH38P13/nib + rsync -a --progress /hive/data/genomes/grcH38P13/{grcH38P13.2bit,chrom.sizes} \ + /hive/data/staging/data/grcH38P13/ + + +##############################################################################